The Triad Primitive
Scenarios
Three parallel instances of the same tool, pre-seeded with different assumptions. Three pillars, one forecast shape — so you can see the spread in a glance instead of clicking back and forth.
Three Runway Zero columns running in parallel. Current trajectory, optimized trajectory, revenue-zero crisis. Pre-seeded, persistent.
Three Task Switching Tax columns — 8 contexts, 3 contexts, single-threaded. Pre-seeded, persistent.
Three Minimum Viable Rate columns at 45% / 60% / 75% utilization. Same target need, different hourly answer.
How the triad primitive works
Each pillar is a namespaced tool instance. On first visit the triad writes pre-seeded state into
separate localStorage keys
(tool__baseline,
__optimized,
__crisis). After that, every
edit persists per-pillar.
Use a triad when the important question is not the single best estimate but the spread between possible outcomes. If the crisis column and optimized column are close together, the decision is probably stable. If they are far apart, the next useful task is to identify which assumption creates the gap.